The guiding question is whether early intervention can improve long term outcomes for psychotic disorders in the U.S. We hypothesize that a specialized multi-component intervention, delivered in a public mental health clinic will be more effective than usual care and also more cost effective. Despite evidence supporting early intervention services abroad, there remains significant uncertainty about its feasibility and cost-effectiveness in the U.S. healthcare system. Data from this study will be critical in informing both service reform and health policy targeting serious mental illness in this country. The core design feature is a pragmatic, randomized controlled trial that seeks to compare specialized multi-component or 'STEP'care to treatment as usual in a clinically meaningful population of individuals who are early in the course of a psychotic disorder. Outcomes will be followed in the domains of clinical status, functional status and costs. Rates of hospitalization over the three year study period will serve as the primary outcome. Outcomes will be measured in three broad domains: 1) clinical status: including hospitalization rate and relapse, suicidality and symptom scales;2) functional status: including ability to work or attend age-appropriate schooling, social functioning and quality of life and 3) costs: broadly measured to include costs of treatment and social and forensic services used. The two interventions being compared by random assignment are (a) care delivered by a specialized team that includes the components of medication prescription, cognitive behavioral therapy groups, family psycho-education and individualized case management that is focused on vocational and educational rehabilitation versus (b) referral to treatment as usual in the community. For dichotomous data, the analysis will employ Fisher's exact tests and logistic regression. The randomized groups will be compared at baseline and significant differences will be entered into the logistic regression analysis. Time to event data, such a time to first new hospitalization after randomization, will be analyzed using Cox regression models. Continuous measures such will be analyzed using mixed regression repeated measures models. The basic method of determining cost-effectiveness will be to collect information on the incidence of cost-related events (service utilization, incarceration, etc.) using established and comprehensive instruments and then to multiply these events by the estimated mean unit cost for each event. We will estimate the economic costs, or the actual costs of resources used, rather than budget costs or fees. Psychotic illnesses are common, disabling and costly under usual care in the U.S. The provision of treatment that is comprehensive and delivered early after psychosis onset has the potential to improve long term outcomes in a cost-effective manner. The Specialized Treatment Early in Psychosis (STEP) project is designed to test the hypothesis that such early care is cost-effective in comparison to usual models of care and will inform the care of individuals who suffer from these common illnesses.